Sweitzer Sara J, Byrd-Williams Courtney E, Ranjit Nalini, Romo-Palafox Maria Jose, Briley Margaret E, Roberts-Gray Cynthia R, Hoelscher Deanna M
J Acad Nutr Diet. 2015 Aug;115(8):1249-59. doi: 10.1016/j.jand.2015.03.012. Epub 2015 May 12.
As early childhood education (ECE) centers become a more common setting for nutrition interventions, a variety of data collection methods are required, based on the center foodservice. ECE centers that require parents to send in meals and/or snacks from home present a unique challenge for accurate nutrition estimation and data collection. We present an observational methodology for recording the contents and temperature of preschool-aged children's lunchboxes and data to support a 2-day vs a 3-day collection period.
Lunchbox observers were trained in visual estimation of foods based on Child and Adult Care Food Program and MyPlate servings and household recommended measures. Trainees weighed and measured foods commonly found in preschool-aged children's lunchboxes and practiced recording accurate descriptions and food temperatures. Training included test assessments of whole-grain bread products, mixed dishes such as macaroni and cheese, and a variety of sandwich preparations. Validity of the estimation method was tested by comparing estimated to actual amounts for several distinct food types. Reliability was assessed by computing the intraclass correlation coefficient for each observer as well as an interrater reliability coefficient across observers. To compare 2- and 3-day observations, 2 of the 3 days of observations were randomly selected for each child and analyzed as a separate dataset. Linear model estimated mean and standard error of whole grains, fruits and vegetables, and amounts of energy, carbohydrates, protein, total fat, saturated fat, dietary fiber, thiamin, riboflavin, niacin, vitamins A and C, calcium, iron, sodium, and dietary fiber per lunch were compared across the 2- and 3-day observation datasets.
The mean estimated amounts across 11 observers were statistically indistinguishable from the measured portion size for each of the 41 test foods, implying that the visual estimation measurement method was valid: intraobserver intraclass correlation coefficients ranged from 0.951 (95% CI 0.91 to 0.97) to 1.0. Across observers, the interrater reliability correlation coefficient was estimated at 0.979 (95% CI 0.957 to 0.993). Comparison of servings of fruits, vegetables, and whole grains showed no significant differences for serving size or mean energy and nutrient content between 2- and 3-day lunch observations.
The methodology is a valid and reliable option for use in research and practice that requires observing and assessing the contents and portion sizes of food items in preschool-aged children's lunchboxes in an ECE setting. The use of visual observation and estimation with Child and Adult Care Food Program and MyPlate serving sizes and household measures over 2 random days of data collection enables food handling to be minimized while obtaining an accurate record of the variety and quantities of foods that young children are exposed to at lunch time.
随着幼儿教育（ECE）中心成为营养干预更常见的场所，基于中心餐饮服务，需要多种数据收集方法。要求家长从家里送来餐食和/或零食的幼儿教育中心在准确的营养评估和数据收集方面面临独特挑战。我们提出一种观察方法，用于记录学龄前儿童午餐盒的内容物和温度以及数据，以支持为期2天与3天的收集期。
午餐盒观察员接受了基于儿童和成人护理食品计划、“我的餐盘”份量以及家庭推荐量的食物视觉评估培训。学员对学龄前儿童午餐盒中常见的食物进行称重和测量，并练习记录准确的描述和食物温度。培训包括对全麦面包产品、通心粉和奶酪等混合菜肴以及各种三明治制作的测试评估。通过比较几种不同食物类型的估计量与实际量来测试估计方法的有效性。通过计算每个观察员的组内相关系数以及观察员之间的评分者间可靠性系数来评估可靠性。为了比较2天和3天的观察结果，为每个孩子随机选择3天观察中的2天，并作为单独的数据集进行分析。比较了2天和3天观察数据集之间每餐的全谷物、水果和蔬菜的估计均值和标准误差，以及能量、碳水化合物、蛋白质、总脂肪、饱和脂肪、膳食纤维、硫胺素、核黄素、烟酸、维生素A和C、钙、铁、钠和膳食纤维的含量。
11名观察员的平均估计量与41种测试食物中每种食物的测量份量在统计学上无显著差异，这意味着视觉估计测量方法是有效的：观察员内组内相关系数范围从0.951（95%CI 0.91至0.97）到1.0。在所有观察员中，评分者间可靠性相关系数估计为0.979（95%CI 0.957至0.993）。水果、蔬菜和全谷物份量的比较表明，2天和3天午餐观察之间在份量大小或平均能量和营养成分方面没有显著差异。
该方法对于在幼儿教育环境中观察和评估学龄前儿童午餐盒中食物项目的内容物和份量大小的研究和实践是一种有效且可靠的选择。在随机的2天数据收集中使用基于儿童和成人护理食品计划、“我的餐盘”份量以及家庭量度的视觉观察和估计，能够在尽量减少食物处理的同时，准确记录幼儿午餐时接触到的食物种类和数量。
